Updated 5-26-97

ISSN 1068-9109

The Goleta Publisher
Dana Trout editor

December 1996 Contents
Master Table of Contents

CD-ROM Recorders
by Dana Trout, GDTPUG

I did it. I bought a CD-ROM recorder (CDR). The prices have come down enough, and the capabilities have improved enough that I felt I could not afford to be without one. And, of course, I have several clients who are now looking to publish their own CD-ROMs, so it would be a good idea for me to know all about it before charging them for my advice. There's precious little worse than dry-labbing when a client's business is at stake.

What I bought was a Philips CD2000 2x recorder/4x player with Corel CD Creator version 2 and Astarte Toast version 3. Corel CD Creator (CDC) comes with a multi-platform manual and CD, and works on Win31, Win95, WinNT, and the Mac. Toast is only for the Mac. The recorder and CDC are being sold now as a bundle for under $600 (external) and Philips includes a $100 rebate. (You have to be a little bit careful about this, though, because they give this rebate only for “retail packaged” drives, not oem. Some dealers are buying the oem version, putting it in their own case, and then selling it and telling buyers they'll get a rebate, when, in fact they won't.)

There are several CDRs with good reputations: the Yamaha 100/102, the Ricoh, and the Philips. I chose the Philips because it has a flash ROM and they post updates on their bulletin board and web site, while Yamaha has a soldered-in EPROM which means I would have to send the drive to them for any updates. Neither supports the new packet recording standard (which itself is in the process of being finalized), so if I later decide that packet recording is important to me, the Philips should be easier, cheaper, and faster to upgrade. Oh yes, the Philips uses a tray while the Yamaha uses a caddy. I plan use the drive only for recording, so for me it is easier to use the tray.

So what are the issues relating to CDR? As usual, that depends on what you want to do. The most common uses of a CDR include the following:

Let's look at each of these uses separately.

Archival Backup Device
CDR has both strengths and weaknesses for this use. It is significantly more stable than tape, and I imagine that drives capable of reading the present discs will exist long after drives capable of reading the present generation of tapes. Tapes really should be retensioned and cleaned periodically, and the data transferred to new tapes within five years or so (the binders holding the magnetic medium to the tape backing does fail over time -- how well they last depends on the quality and the care given to the tapes). Tapes should not be placed in a vault and then forgotten.

However, the current CD capacity of only 660MB is a problem in the era of consumer computers with 1.6GB drives. So you really need to decide what it is you want to archive. My personal feeling is that there is little need to back up the applications (each with their multi-hundred MB of files) or the operating system: after all, you still have the CDs they came on. However, the stuff you create with the applications is what needs archiving. Fortunately, most of us create only megabytes of files a week with our word-processors, spreadsheets, and so on. The CD is OK for this volume of data.

Those who are doing live video will find the CD to be much too small for archiving: there's just too much data. Live video producers are complaining vociferously about the 2GB limit in Win95 (hence the new FAT32 file structure) because they can't fit more than a few minutes in 2GB. The CD is not for them as an archiving tool (but can be just the ticket for distribution -- we'll get into this later). But for multigigabyte storage the only solution is tape.

Other possibilities for archiving are the Iomega Zip and Jaz, and the Syquest Easy 230 and 1.3G. However, when you look at the cost per MB they all come out about the same, which is about five to seven times the cost of CDR. Strangely enough, the cost of CDR media is closer to the cost of tape than to other removable cartridge systems, but it is much longer-lived than either.

The problem with getting such a powerful way of archiving is that you suddenly realize that you can either record everything in a hodge-podge with minimally labelled discs, or try to be more organized. And the problem with organization is that you have to do it, it takes time, and there's too many other things to do right now. The best I can suggest is organize your archiving process as much as you can without distracting yourself from what you really need to do.

In my case I have Win95 on a 800MB disk (and Win31 on a 1GB, go figure). Anyway, the Win95 disk was bursting at the seams (I even had to remove PageMaker for a while) with stuff I wanted to keep available, but it didn't really have to be mounted or on-line all the time. The CDR was the perfect solution: most of my computers have CD players, I could put all that detritus on several CDs and slip them in when I needed them. This gave me not only archiving but near-line storage as well.

Near-Line Storage
This term refers to data that isn't permanently stored on the computer, but is readily accessible. We are all used to using floppies for this purpose, but with the size of Word and PowerPoint files these days, floppies don't hack it. The Zip has become wildly popular as the stand-in for a bigger floppy, but its cartridge costs twice as much as a CDR blank while holding only a sixth of the data. Also, most computers now have CD readers, but relatively few have Zip (or other removable cartridge) drives. That means if you are looking at the total cost of the system, you will have to compare the price of adding a Zip drive to every computer (plus media costs) to the cost of $6-8 per CDR blank for each computer and one CD recorder.

Low-Cost Shipping of Lots of Data
Think about it: a $6-8 CD plus tyvek sleeve plus two sheets of cardboard and an envelope compared to a $95 Jaz cartridge plus large padded envelope for its carton. And when the data gets to the destination what is the relative likelihood that the computer there has a Jaz drive as compared to a CD reader?

Low-Cost CD Master for Replication
CD replication houses do accept files on other media, such as Jaz cartridges, DAT tapes, and so on. But then they have to build a CD disk image before making the master (and they charge you for this service). The master they build may not be what you actually wanted: especially in the case of the Macintosh you get to “build the desktop” which means you get to choose the position of the icons and windows. If you make your own CD master you can lay it out exactly the way you want, and you can test it for performance. If you need to make changes, you just burn another one for $8 instead of finding yourself stuck with 5,000 copies that really aren't quite right. When you've finally come up with the CD layout you want, then you send that to the replicator and that's what you get thousands of. In this case the low cost primarily refers to the low cost of an error -- you get to find and correct that error before it becomes a high-cost item.

CD Recording
OK, so you think recording on a CDR is for you: your data fits well into the approximately 660MB size, you like the fact that most computers now have a drive that can read your recorded data (not true of tapes or removable cartridges), and you like the fact that the data is readily accessible (getting a particular file on a CD takes less than a second, while listening to a tape whine for a minute or so can be wearing). What are the “gotchas”?

CDR Recording Gotchas
First I'll need to back up and tell you a little bit about CD technology. Then you'll be better able to understand the issues.

A CD is not laid out as concentric tracks like a hard or floppy disk: it uses a single spiral groove, much like a record (but unlike a record, the groove starts near the center and ends near the edge). The groove is broken up into blocks with approximately 2KB of data each (the exact amount depends on the kind of data: sound, digital data, and so on). Each block is numbered, and there is a Table of Contents (TOC) which lists the contents and the corresponding block numbers. There is also a lead-in at the beginning and a lead-out at the end.

So why do we care about how the disc is laid out? Because we have some choices about how we record which affect the readability of the disc on various CD drives.

Our choices for the overall structure of a data disc are Single Session, Multi-Session, and Multi-Volume. All CD readers can read a single-session CD. Most drives made within the past three years can read multi-session (if your drive can read all sessions on Kodak PhotoCD discs, it can read multi-session). Most multi-session capable CD readers can read multi-volume but need a special software driver.

Multi-volume is the easiest to record. Say you want to archive all of your My Documents directory once a week. The first week you record (for instance) volume 961104, the next 961111, and so on. Each volume is placed on the CD as a complete unit with its own TOC and data. Although this is the easiest to record, it is the hardest to play on other computers. Yours will have the needed software driver (it comes with the CDR software) but most other computers don't have it -- they never needed it. Another point is that you can see only one volume at a time. For instance, if you selected the 961104 volume, you would not be able to read the files in the 961111 volume. It's trivial to switch volumes, but if you are using a scanning tool like FindFile, you will have to manually switch volumes and rerun it until you find what you are looking for.

Multi-session differs from multi-volume in that there is only one TOC. When you record a new session, it is compared with previous sessions and a new TOC which refers to all files for all sessions is created. The old TOCs are ignored -- they just become bits of inaccessible data debris. If you use Corel's CD Creator to do multi-session recording you can also tell it to leave out files from previous sessions in the new TOC. This way you can “delete” files on the CD (they are still there, but nothing points to them so they are relatively inaccessible). And, of course, this is the way you can replace a file with a newer version. From the standpoint of reading the disc, most players and software made in the past three years will work just fine.

Multi-session is compatible with more players and operating systems than multi-volume. However it chews up more of the CD with its ever-growing copies of the TOC. The multi-volume gives you true traceable archiving: files cannot be deleted or overwritten, which means that multi-volume is a true snapshot at a particular point in time. Whether this is a feature or a curse depends on what you need.

Both multi-session and multi-volume begin each session (volume) with a lead-in and end with a lead-out. This wastes some storage capacity and may be a concern if the amount of data you need to record is nearly the capacity of the CD.

Single-session can be read by anything that purports to be a CD player (including your audio CD player, but you will probably blow out either the amplifier or the speakers, so don't do it). The downside is that you get only the one session on the disc: if you have only 20MB of data you've just wasted the remaining 600MB+. That's a pretty hefty price to pay to assure backward compatibility for drives over three years old. Ten such discs will pay for a new CD reader. But there are times when it makes sense: if you have to send out weekly summaries to 20 clients the concept of multisession loses its appeal. Sure, you could ask the clients to send back the previous disc, but when you balance the cost of shipping and handling (plus the inevitable scratches due to mishandling) against the $8 cost of new media you find the decision is quite easy.

The TAO of Speed
TAO stands for (in this case) Track At Once. It is pitted against DAO (Disc At Once) and SAO (Session At Once). A session (even the one session of a single-session disc) can be composed of just one track (which is what you get in all commercially-produced data CDs) or multiple tracks. The reason you'll probably wind up recording multiple-track sessions is that TAO puts the least demands on your computer.

Speed is an important issue because once the CDR fires up its laser and starts recording down that spiral, you've got to feed it the data it needs when it needs it all the way to the logical end of the recording. It first writes the lead-in and TOC, then the files, and finishes with the lead-out. If the recorder runs out of data before it reaches the end the TOC points to data that really isn't there -- you've just made an expensive and not very effective coaster.

Writing DAO (disc at once) is the hardest to do because you've got to have all 660MB data right at hand. SAO is easier because you need only a session's worth at a time. And TAO is the easiest of all because you can make the track length a convenient size.

Lack of Speed Kills
You don't need blazing speed to feed the CDR, but you have to feed it consistently. The CDR needs only 150KB/sec for single speed recording, 300KB/sec for double, or 600KB/sec for quad. (Sound requires slightly higher data rates: 170KB/sec for single, and proportionately more for double and quad speed.) These data rates really aren't very fast, the rub is that you can't pause and do something else. So of course you turn off screen savers, background tasks, and anything else that might interrupt the smooth flow of data. If you are on a network and one of your drives or printers is shared, remember to logout during the recording.

You will have to look at your setup to make sure you can feed the recorder. For instance, I have one computer with a 486dx4/120MHz but with an Adaptec 1520 SCSI card to connect to the CDR. The processor is fast but the SCSI interface is slow (not bus mastering). The combination works even at 2x recording speed because the processor speed makes up for the turgid interface. And it helps that the hard drives I am recording are not on that same SCSI bus (they are on an IDE interface).

Another computer is a 486dx2/66 running only half the speed of the first. But it has an Adaptec 1542 bus-mastering SCSI card which connects not only to the CDR but all the other disk drives. The card's enhanced throughput makes up for the moderate processor speed and the fact that everything runs through that one card. And this combination works fine too for 2x recordings.

I know that neither of these setups would work at 4x speed unless I built an actual image of the CD on a hard drive (which means setting aside a 700MB scratch area which I was unwilling to do), so I didn't pay twice as much for a 4x recorder that I would use most of the time at 2x. Also, I feel that in a year or two the prices will drop because of the introduction of DVD recorders (which will record eight times as much information on the same-sized disc). Actually, what probably will happen is the prices stay about the same but the only recorders will be DVDs, but the conclusion is the same: why spend that much premium for so little real advantage?

December 1996 Contents
Master Table of Contents

Corel CD Creator Version 2
review by Dana Trout, GDTPUG

Corel CD Creator (CDC) is a popular application program for controlling CD-ROM recorders, and is often bundled with them. It comes with a multi-platform manual and CD, and works on Win31, Win95, WinNT, and the Macintosh.

At first I was frustrated by Corel's CDC manual. What bothered me was that there was no information about system requirements (RAM, how much hard disk would be needed, etc), nor, for that matter, any information about software installation. I tried looking things up in the index, but it referred me to page A-9. But I couldn't find any Appendix A. When I looked at other entries in the index, I noticed that they all started with an A, B, or C. Finally the light dawned: they were referring to different sections of the manual, not appendices. It still seems unduly complicated, though, because section A starts on page A-1 (natch), B on page B-43 (huh?), and C on page C-83 (double huh?). Canadians sure have some funny ideas about numbering sometimes.

The manual is pretty superficial, but excuses itself by telling you to look at the help files. That would be OK except for the fact that the help files are pretty superficial too. The most irritating response to clicking Help in any of the applications is The topic does not exist. Contact your application vendor for an updated Help file. (129) But for the most part the software performs OK.

Note that many of these applications are not really full-fledged. For instance, even though PCD Creator lets you copy images from Kodak PhotoCDs, the resulting disc cannot be played on a PhotoCD player. You can, however, read the resulting disc as you would any other Kodak PhotoCD on a computer. Also, PCD Creator has no image manipulation tools -- it merely lets you drag and drop images from the source to the destination. Within its limits, it is easy to use and effective.

The Sound Editor is better than the Windows Recorder, but no one would mistake it for a full-featured or professional tool. It allows you to load, record, and save WAV files at various sampling rates (8, 11, 22, and 44KHz), bit resolution (8 or 16), and as either stereo or mono. You can apply several effects: Amplify, DC Offset, Echo, Fade, Flange, Graphic Equalization, Invert, Pitch Shift, and Pop Removal. The basic tools are there but some of the more sophisticated capabilities are not. For instance, the purpose of DC Offset is to remove a dc offset which will cause increased noise or hiss when other effects are applied. The problem is that you have to enter the amount of offset to apply: what would make far more sense is to tell the computer to figure out what the dc offset is and remove it. The real problem is that eyeballing the waveform and guessing at the dc offset is hard to do accurately.

The main tool, and fortunately the one that they thought out the best, is CDC itself. This tool opens with a wizard which leads you through the decision steps needed to create or append to a CD. Of course some of the questions are terse, but generally they give you enough information to make a reasonable decision. The only thing I've done so far is create data CDs. The first question I am asked is whether I want to make an ISO-9660 or Joliet disc. For ISO I am told “The ISO 9660 is the most widely supported standart(sic) for CD-ROM discs. Discs can be used on the platforms such as DOS, OS/2, Windows, UNIX and Macintosh.” For Joliet I see “Discs created using the Joilet(sic) file system can be read only under Windows 95. File and directory names can be at most 64 characters long.” Fortunately the manual is more detailed. There really are several more issues here: one is that some of the characters allowed (such as - and !) in DOS file names are not allowed under ISO 9660. Another is that ISO 9660 allows no more than eight levels of directory nesting, while DOS doesn't care. And finally, directories are not allowed to have extensions under ISO 9660 (Microsoft is famous for creating directories with extensions for its own applications). So even if you are using only short file names with Win95, your best choice is probably the Joliet file system unless you can guarantee that your directory structure will not exceed eight levels, no directory has an extension, and you use only the characters A to Z, 0 to 9, and underscore. If you are using Win3.1 you can tell CDC to relax some of the 9660 restrictions, with the upshot being that these discs will play on Windows computers but not on Unix or Mac.

After you have answered all the wizard's questions you can tell CDC to first run a simulation, then, if all goes well, actually write ("burn") the disc. In doing the simulation CDC performs every step, including telling the recorder where to position the head and feeding it the data, that it does for writing, with the only difference being that it does not tell the recorder to turn on the laser to high-power (write) mode. If CDC can make it all the way through the simulation then it will make it through the write OK as long as the computer is not interrupted. (You did remember to disable the fax, and modem answer, and screensaver, etc. didn't you?) The way it copies the files to the CD is interesting: first it looks to see how much hard disk space it can use for a scratch area. Then it goes through the list of files to be copied and pairs the longest with the shortest, second longest with second shortest, and so on down the list. The purpose of this little exercise is to alleviate the problem of getting a zillion short files in a row (each with a time-consuming directory lookup) and running out of time to feed the recorder. Then it decides how many tracks to make this session (or volume, as the case may be), based on the amount of hard disk scratch space, the recording speed, and the speed of the system. Small scratch area, slow system, high recording speed all make for more tracks, because each of these things means that the system will have a hard time keeping up with the recorder. And more tracks means each is shorter, so the recorder will come to the end of the track before CDC runs dry on data.

CDC uses the hard disk scratch area to note where each file is actually kept, so it does not have to rattle through the directory structure while burning the disc. This helps a lot, especially in directories which have lots (hundreds to thousands) of files, because DOS (and Windows) are real stupid about how they look for a file: They look at the first entry, and if it doesn't match, they look at the second, and so on. The file names are not indexed in any way so there is no speedy way of finding a specific one. If you are looking for the thousandth file DOS finds it on the thousandth try. And if you need the one right after it, DOS has to go through all thousand again and one more.

CDC's approach really does help, but it isn't foolproof. First, of course, is if you have a lot of very short files it still takes a lot of time to retrieve them compared to the amount of data you get. The second is that if you do anything between the time CDC builds this index and the time it actually uses it, you're not going to be too happy. CDC does include a tool called “Validate” just to cover this case.

Recently CDC has been purchased by Adaptec and they are busily melding its features with their own product, Easy-CD. Everyone is looking forward to getting the best features of each, but the finished product is not expected until early 1997. In the meantime there are several excellent web sites devoted to CDR (and CDC) issues (such as http://www.cd-info.com), including Adaptec's own (http://websvr1.adaptec.com/support/cdrec/faqindex.html).

In conclusion, after one gets used to the fact that some information is in the manual and other necessary information is in the help files, and some statements are ambiguous and all you can do is try it out and see what happens, CDC is an acceptable program for writing files to a CD recorder. Unfortunately, Corel in its inimitable fashion opted for putting on more barely usable bells and whistles (e.g the sound editor and PCD Creator) than for a good, solid, complete basic design and manual. So far all the CDs I have recorded with CDC have worked flawlessly, so I shouldn't complain too much about the software. But it would help if the manual were more lucid.

One last observation: the price of any CDR software is several hundred dollars if purchased separately, but less than $100 if purchased with the recorder. If you are interested (as I was) in getting several different CDR programs (I got both CDC and Toast) buy them all at the same time you buy the recorder.

December 1996 Contents
Master Table of Contents

The QuarkXPress Book 3.3
by David Blatner and Bob Weibel
review by Christine Nolt, GDTPUG

Anyone who has yearned to design with type and images, to place and tweak with precision, and to output files in a variety of ways, will find QuarkXPress for Windows an indispensable program. It is simple enough in its design to let a novice get up and running quickly, yet powerful enough to handle difficult typographical challenges. As a word “processing” program it doesn't do many routines needed in an office environment (e.g. automatic tables and graphs or indexing) without the additional purchase of plug-in x-tensions. But weighed against reliable color output, extreme design flexibility and intelligent and creative user it's a valuable tool and an enjoyable program to learn.

The book to be reviewed in this article, The QuarkXPress Book by David Blatner and Bob Weibel, is published by Peachpit Press, contains 662 pages including a seven part appendix and index. Also included is a tear-out reference card and an offer for a $10.00 Goodies Disk. It sells for $29.95.

The authors of the Quark book write with clarity and with very little repetition. They understand the program (which is required of any author of this type of book) but more importantly they understand the process of type design. They approach the explanation of Quark, not as sales material discussing it feature by feature, but rather from an understanding of how Quark is applied in working situations. The book is not a tutorial, per se, but provides comprehensive descriptions of many features of the program. There are ten main chapters beginning with the “Basics”: “Document Construction”, “Word Processing”, “Type and Typography” (100 plus pages of valuable resource information for any desktop designer), “Copy Flow” (using styles to format text) “Pictures”, “Where Text Meets Graphics”, “Image Modification”, “Color” and “Printing”.

In chapter nine on “Color”, the reader will find explanations of the differences between process color and spot color, as well as the differences between various color models that Quark handles (RGB, CMYK, HSB, Pantone, Focotone and Truematch). Suggestions are given on creating a color list and how to go about naming your colors. The chapter like all the others in the book includes “Tips”. These often present observations on a procedure being discussed or relevant background information. For instance, the tip on page 395 states, Color Tricks the Eye. Placing a colored object next to another differently colored object makes both colors look different than if you had just one color alone. Similarly, a color can look totally different if you place it on a black background rather than a white one.

The tip concludes by offering suggestions on how to get a truer sense of color from swatch book samples. The chapter also gives a solid discussion on the topic of trapping with diagrams, clarifying the information presented.

I found the presentation of information concise and well-written and consider it a valuable reference book. The excellent index and detailed table of contents make finding an answer to a problem or question a snap. As a reference book it is both simple enough to help a first time user (both in language and conceptual presentation) while offering detailed procedural information to the intermediate or advanced user.

The QuarkXPress Book 3.3
by David Blatner and Bob Weibel
Peachpit Press (800) 283-9444 http://www.peachpit.com
$29.95 ISBN 1-56609-135-7

December 1996 Contents
Master Table of Contents

They're Only Words
by Dana Trout, GDTPUG

Typos are so Freudian. One email I got referred to an “internet hosing service”.

Another was an ad for a lost Queensland Healer. Yes, that is a very valuable property, much more than just a pet: anything that can promote healing after the recent political season definitely needs to be found, not lost.

One more was a reference to a person being the lynchpin of an organization -- I hope they meant linchpin, “a locking pin inserted in the end of a shaft, as in an axle to keep a wheel from falling off” according to the American Heritage Dictionary. I presume a lynchpin is the peg you attach the hangman's noose to -- does that mean this person was in charge of massive layoffs?

December 1996 Contents
Master Table of Contents


Back to The Goleta Publisher home page
Goleta DeskTop Publishing UG
P.O. Box 8450
Goleta, CA 93118-8450
(805) 685-7937 voice or fax

Copyright Notice: All material published in the Goleta Publisher is copyrighted. Responsibility for, and ownership of, copyright remains with the author of each article, or with the publication from which the article was reprinted.
email to: Dana Trout, Web author